^{page 104 note 1} A 46, B 467, r 259, 381, A 584 (twice, or imperfect), N 339, 543, 392, II 280, 340, P 596, T 61, 444, Φ 389, 407; γàρ, N 608. (six instances) is neglected, for a reason which will be given later. In the similes, δΕ—O 581, X 94; γε (= ‘ both ’)—T 168.

^{page 105 note 1} If A 175 = P63 are instances at all, they are forced; cannot be satisfactorily brought normal in the strong caesura (one overlength in into the verse.

^{page 105 note 2} The statistics are disturbed by A (48, 309, 311, 436, 460); in all points which I have been able to tabulate, the book is purely Odyssean from end to end.

^{page 105 note 3} By ‘overlengtb’ I mean more consonants than are necessary to give length—e.g., Overlengths are normal in the strong caesura (one overlength in every three occurrences of the break); but the ninth-hemimeral seems to reject such scansions, especially the long closed syllable followed by another consonant (type -), which are perhaps as infrequent as 1 in 30 occurrences of the break (Cl. Q., April, 1908, pp. 94 sqq.).

^{page 106 note 1} The short syllabic augment (e.g. ) is alone considered here; the long syllabic (e.g. ) is discussed later

^{page 107 note 1} Reckoning and of course.The proportion of such augments is lower in the Odyssey (not quite 1 in 4), and much lower in B2 etc.(about 1 in 6).

^{page 107 note 2} is not counted here (nineteen instances); but it probably began with two consonants, though the first of these (poor ghost!) is exorcised in Ψ 679.

^{page 107 note 3} By a dovetail I mean a short final vowel (about 1 in 6). scanned long before two initial consonants, as in

^{page 108 note 1} Of these eleven are of type arrive, and four are of type It shows how real was the restriction of overlength, that words like and should so often be placed after the strong caesura, where they give the same difficulties as those presented by an open antispast— must be elided, or dovetailed, or lengthened with -v; has not the third resource. Aorists like have nothing to motive this difficult scansion, and only 12 out of 126 are placed after the strong caesura, while of the fourteen in type , none are so placed. No doubt the ratio in type (1: 10½) would have been even lower, except for the influence of type (i:5).

^{page 108 note 2} For the rest of narrative, add—12 442; θ 118, x 181; α 127, Ψ 478, ζ 211, λ 334, v 2, ξ 420, ρ 29; imperfect ζ 226. The elisions , are not counted here, but are considered, as the phrases actually stand, metrically certain

^{page 110 note 1} Except in 0 280.

^{page 110 note 2} In the imperfect—M 254, Φ 571; the imperfect is always less law-abiding than the aorist.

^{page 111 note 1} Also found in a simile, F 35; but the passage is not good evidence. The similes in T, like the surprises in the Wrong Box, ‘might have shaken the reason of Immanuel Kant.’

^{page 112 note 1} Not wholly unlike are (sixth foot )

^{page 112 note 2} (A 112) seems still unsolved; are all difficult.

^{page 112 note 3} By false length I mean scansions such as would be irregular, but not unlike F.

^{page 113 note 1} .

^{page 115 note 1} The forms , and the imperfect of are omitted.

^{page 117 note 1} 1 177, 484, 541, K 97, 148, 220, 310, 397, 39 T'λ 471, 471, 546, 615, μ 204, 232; imperfect μ 332

^{page 117 note 2} I have not tried the Odyssey; in reckonings for the augment, dies citius deficit quam nomina.

^{page 118 note 1} γ 342, 395, 461, e 279, 295, 17 184, 228, p 326 458, δ387, 427, Π 273, ω 513; imperfect e 196 θ 470, v 69.

^{page 118 note 2} B 317, 326, I 492, 545, O 191, 607, 631, δ 95, e 223, η 268, θ 155, 0 367, p 483, w 65; K 242, 196, μ 13, 364. Imperfect–M 165, X 298, f 171, v 131, 206, 146, to 24, 284.

^{page 119 note 1} These forms are not absolute, like but is much less convenient than and both and have serious objections.

^{page 120 note 1} The same thing seems to hold true of the imperfect; but the tense is a mere labyrinth of perplexities, and only complete enumeration could satisfy the reader. To compare the Iliad and Odyssey, the ratio of preferentials must not be taken alone, but must be combined with the ratio of augmentation.